ALBERTL. HENNEAND ROBERT C. ARNOLD
758
Vol.
TABLE I11 Sample
Total time of reaction, hours
First run
c
cis-
Dichloro, g.
[a]of cis-
filtrate
+0.06
(-0.09) - .13 - -18
Original
A B C D E
1 5 12 24 48
1.92 3.64
0.96 0.88 0.24
c
[a1 of
dichloro
+ .05 + .05 + .02 + .04
of [Co en2 ( N O I ) ~+. ] The best yield of optically active dinitro product that was ever obtained was 40%. It is apparently possible to convert one form of the [Co en2 ti-tart] Br to the tiiriitro salt without affecting the othcr form. When the reaction is allowed to proceed at room teinpcmture, the precipitation of calcium tartrate goes 011 slowly but steadily until about half is precipitated; the reaction then stops unless the temperature is raised.
Summary The presence of two or three optically active chelate groups in an octahedral complex tends to fix a definite configuration upon the complex ion as a whole, and in some cases the number of stereoisomers which can be isolated is limited to a small fraction of those theoretically possible. No such effect has been observed in complex ions containing only one asymmetric chelate group. It is now shown that while both the D and L forms of dextro-tartrato-bis-ethylenediamine cobaltic ion, [Co enz d-tart ] +, form when dextro-tartartic acid reacts with [Co enz Coal+, they differ greatly in reactivity. When the mixture of the two is
-
cis-
Dichloro, g.
Second run[.I of cisdichloro
io
[a1 of
filt.
(-0.09) +0.01 .00
1.13 4.30 0.75 0.80 0.34
.26 .34 .40
.00
.01 .00
-0.12
-
-
-
.19 .24 .33 .40
shaken with ethylenediamine at room temperature, part of the material reacts within two hours, giving dextro-[Co ena]+++. The remainder does not react even in twelve hours and can be recovered. At SOo, a io% yield of dwtro-[Co e%]+++ is obtained, and very little of the original material can be recovered. Evidently the less reactive form changes to the more reactive as the latter is used up. The reaction with hydrochloric acid seems to follow a similar course, and gives a large yield of dextro-rotatory cis-[Co enz Clz]C1. Reaction with calcium nitrite gives similar results; a t room temperatures or slightly above, about half of the tartrato salt reacts, yielding dextro-[Co en2 (NO&]+. The remainder will not react at room temperatures. At higher temperatures the product is racemized. These are evidently the first examples of asymmetric synthesis in the field of inorganic complex ions. URBANA, ILLINOIS N E W ORLEANS,
[CONTRIBUTION FROM THE DEPARTMENT OF CHEMISrRY AT
LA.
THEOHIO
RECEIVED JUNE 21, 1947
STATE UNIVERSITY]
Hydrofluorination in the Presence of Boron Fluoride BY ALRERTL. HENNEAND ROBERTC. ARNOLD The addition of hydrogen fluoride to an olefinic hydrocarbon occurs with great easel ; addition also takes place easily with asymmetrically halogenated olefins such as C H z d H X , RCH=CX2 or RCX=CHZ,~while the presence of the vinylic halogens on both sides of the double bond impedes or prevents We have now observed that when the addition is carried out in the presence of small quantities of boron fluoride the reaction is greatly hastened, and even occurs where it had failed in the absence of boron fluoride. Besides plain hydrofluorination, an alternation of HI? additions and HX reinovals is observed, so that RCII--CXz gives not only RCH2CX2F, but also RCH=CFX, RCHzCXFz, RCH==CFz and RCHzCFa in sequence. The presence of the boron fluoride increases the number of these steps and the extent to which they occur; the net effect is to increase the fluorine content of the final product. (1) Grosse and Lind, J . Org. Chew., 8 , 26 (1938). (2) H e m e and Plueddeman, Tiirs J O U R N A L , 66, 1271 (1943), also Renoll, i b i d . , 64, Ill!! (1942)
The use of boron fluoride is particularly advantageous with heavily halogenated olefins, or with olefins bearing a CFa group which slows down the reactivity of their double bond. For more reactive olefins, boron fluoride may not be so desirable because it hastens resinification as well as addition; the economy of each case should be individually considered. The literature statement' that BFa is of no help should be restricted to non-halogenated olefins, where addition is so easy that it does not need any help. Eleven different olefins were treated with hydrogen fluoride in thc molar ratio of 1/2; when UFa was added, it was in the concentratioii shown in grams per mole of olefin in the table of experimental results. The observed improvements can be explained by the formation of a coordinated complex HF-+ BFs, which enhances the percentage of ionic character of the H-F bond; this facilitates the separation of hydrogen as a proton, and makes it more available for addition to the more negative of the
EXPERIMENTAL RESULTS
+ HF-+ CHzClCClzF (A) + CHzClCClFz (B)
CHCl--CClz OC.
60 95 95 95 95 95 95 95 95 120 160
Hours
BFa, g./mole
mole % A
mole % ’ B
24 12 3 12 1 3 12 24 3 20 12
12 3 6 6 12 12 12 12 24 20 0
18 16 19 35 16 20 45 60 38 35 0
0 3 2 9 2 3 10 9 2 22 0
+ I-IF
CHCl=-CHCI 21 25 21 60 11 60 21 60 3 95 3 95 11 95 1 95 3 95 11 160
la 13 0
+
€11‘
+ HF
+ IIF
+
95 150 180 180 I80 95 120
ery
88 92 90 82 93 90 84 86 87 57 95
+
CH2C1CHCIF(A) polymer (B) 3 10 83 17 17 80 26 25 82 26 33 68 10 9 82 21 17 80 20 29 54 19 26 79 24 29 63 0 0 60
(B) 97 87 80
+
CF,ClCHClF (A) polymer (B) 0 0 89 trace trace 96 43 trace 87 0 trace 95
+
CFzClCHClr (A) CFZ=CC12 (B) 33 17 82 0 0 0 89
-+
24
CCl,=CCI. liF CCI~FCHCI~ (A) 0 1:3 30 95 7 7 24 140 15 160 14 24 trace 17 24 180 0 17 0 1EO -+
+ HF
CHCI=CFj 0 5 3 3 0
CFSCCi =CCL
recov-
+
-+
24 24 24 0
mole %
CF3CHC12 (A) polymer 0 0 42 traces 0 traces
4
24 24 0
CFCI=CFCl 95 18 150 12 180 17 180 17 CFCI=CCls 180 17 180 17
4
12 6 12 12 3 6 6
CFz=CCI, 95 10 150 12 150 12
25 95
750
HYDROFLUORINATION IN THE PRESENCE OF BORONFLUORIDE
Feb., 1948
+ HF
18
14
12 17 17 17
24 24 30
-+
90 70
+ CFCI-ICC~~(B) 0 trace 3 20 0
CHiClCF, (A) 0 0
CI:jCH=SClE’ 24 5 24 0
$0 90
+
CF~CHCICCIZF (A) CFJCCI= CClF (B) 0 0 88 9 trace 91 22 7 90 21 10 91 0 0 92
--+
0
82 82 77 87 97
+ HF
-+
65 0
CF.ICH~CCIF~ (A) 0 0
85 85
CH,CH=CHCI
60 100
+ CH3CHsCHClF (A) f CHaCHFCHnCl (B) 2 12 20 56 10 20 60 0
6 4
CH,CCl=CHCl 25 95
+ HF
+ HF
CH&!CIFCHzCl (A) CFzCHzCl (B) 2 35 4 0 5 10
5 4
+
+ CHI76 92
doubly bonded carbons. In support of this interpretation is the opposite fact that hydrogen fluoride addition to alkynes is slowed down by an oxygenated solvent such as ether or acetone3; in this case, complex formation with the solvent involves hydrogen bonding to the unshared electrons of the oxygen atom and obstructs separation as a proton for the first stage of addition to the triple bond. The direction of addition was always that expected from the ionic character of the olefin. The vinylic halides acted as if entirely in their U
resonating form
n
:CH~-C , It i i
so that RR’C===
CR”X lead exclusively to HRR‘C=CR”XF. By induction, a CFa group attached to a doubly bonded carbon caused a polarization opposed to that created by a CH3 group, so that the polarization of propene and trifluoropropene could be represented as CH&H=CH2 and CF3CH=CHz1 re-
+
-
+
-
spectively. The ease of addition was not determined solely by the extent of polarization of the double bond. CH2=CC12 accepted hydrogen fluoride more readily than CHCl=CHCl and C C l d F 2 more than C C l F 4 C l F ; CHsCCl=CClH was more easily hydrofluorinated than CFaCH= CClF, yet the latter compound should be much more polarized than the former due to the combination of the eletronegative character of the CFO group mid the vinylic character of the =CFC1 group. It is proposed that the absolute electron density around the double bond should also be taken into account; if this is done, it is to be expected that all ethylenes bearing a CFs q o u p should be slowed down by the electron drainage away from the double bond; in agreement, HCCl=CC12 was found more easily hydrofluorinated than CF3CCl=CC&. This explanation would be similar to that used for the meta-directing effect of an electronegative group on a benzene ring, and its retarding influence. Experimental Details Reaction Equipment.---Iqor relatively low pressure, an 800-ml. steel vessel fitted as previously described4 was used; for the higher pressures prevailing with more boron fluoride, a 200 cc. Aminco bomb was used, with superpressure fittings. Heating was done by immerging in a hot water-bath, or by strapping to the electrically Heated mantle of a mechanical rocking devise. Reagents.-Commercial grades. (3) Henne and Plueddeman, THIS JOURNAL, 66, 587 (1943). (4) Henne and Plannagan, t b r d , 66, 2362 (1043).
760
R. H. ANDERSONAND D. H. WHEELER
Loading Procedure.-As previously described . 5 . 6 Metering of Boron Fluoride.-A steel cylinder of known capacity was connected to a commercial tank of boron fluoride; the gas was admitted and the pressure created was noted. Next, this loaded steel cylinder was connected to the reaction vessel and the amount of boron fluoride delivered was computed from the observed drop in pressure. For very small additions of boron fluoride, a stream of the gas was passed through weighed liquefied hydrogen fluoride cooled to -70". Treatment of Reaction Products.-Gaseous products were released through a washing train and collected in a Dry Ice cooled receiver. Liquids were poured on cracked ice, neutralized, steam distilled, dried and separated by fractional distillation. In the few cases where the olefin and its addition product boiled practically a t the same temperature, the composition of the distillate was esti( 5 ) Henne and Haeckl, THISJ O U R N A L , 6S, 2692 (1941) (F) Henne and Whnley, ibid , 64, 1187 (1942).
LroI. fO
mated from the index of refraction, but this procedure was avoided as much as possible by using efficient dephlegmators.
summary The addition of hydrogen fluoride to halogenated olefins is accelerated or made possible by using small amounts of boron fluoride as a catalyst. The action is attributed to the formation of a coordinate compound H F -+BFs, in which the H to F bond is weakened; this favors the supply of protons for the first step of the addition to the double bond. The direction and ease of addition are discussed. Eleven olefins have been tried and the results are tabulated. RECEIVED SEPTEMDER 2 2 , 1017
COLUMBUS, OHIO
[CONTRIBUTION FROM GENERAL MILLS,INC., RESEARCH DEPARTMENT]
Polyamide Resins from Dilinoleic Acid and Ethylenediamine. Viscosity Relationships la
Molecular Weight-
BY R. H. ANDERSONAND D. H. WHEELER
Introduction In 1941, Bradley and Johnston'" isolated dilinoleic acid in a pure state as the methyl ester by fractional molecular distillation of the heat-polymerized metbyl esters of the acids from dehydrated castor oil. They prepared polyesters of triethylene glycol with molecular weights of about 4300. Cowan and Wheeler2 prepared polyesters with various glycols having molecular weights up to 20,000 to 30,000. Polyamides have been prepared by Falkenburg, Teeter, Skell and Cowan3 from various diamines and the residual polymeric fatty acids or esters (mixtures of dimer and trimers of linoleic and linolenic acids). The polyamide from ethylenediamine and these mixed polymeric fatty acids was proposed as a new synthetic coating material4 under the name Norelac. The molecular weights of these resins were 3,000 to 5,000, determined by the method of end-groups. The same type of polyamide has been commercially prepared on a pilot plant scaleI5arid is apparently finding some use as a coating for heat-sealing and laminating paper. No description has appeared of polyamides from pure dilinoleic acid and ethylenediamine. The present investigation describes the prepkration and some properties of such linear polyamides having molecular weights from 2,000 to 15,000. Viscosities were determined on solutions in 1: 1 (la) Paper No. 79, Journal Series, General Mills, Inc., Research Department. (1") Bradley and Johnston,'Znd. Eng. Chem., 99, 86 (1941). (2) Cowan a n d Wheeler, THIS JOURNAL, 66, 84 (1944). (3) Falkenburg, 'Teeter, Skell and Cowan, Oil a n d S o a p , 22, 1 4 3 (1945). (4) Cowan, Lewis a n d Falkenburg, i b i d . , 21, 101 (19.14). ( 5 ) C o w a n , Schwab and Falkenburg, M o d e r n P o c k a g i i r g , [SI 11, 103 (1944).
butanol-toluene solvent a t concentratioiis from 0.2 to 9.8 g. per 100 cc. and intrinsic viscosities were calculated. The intrinsic viscosity was found to be related to molecular weight by the formula [SI =
KM~
(1)
Flory aiid Stickney6 have pointed out the fact that a relationship such as equation (1) really relates to wiscosity average (Mv)rather than number average (Mn) molecular weights of non-homogeneous polymers, and that M , is greater than iWn, but less than Mw,the weight average molecular weight. Now, the inolecular weights, as determined by end-group titration in the present study, are actually number average molecular weights. The fact that the determined molecular weights and viscosities of these polyamides agree with equation (1) is interpreted to mean that these polymers, as prepared, have essentially a constant ratio of ;21, to Mn. Flory' has shown on theoretical grounds that for bifunctional condensation polymers, AI, 1 Mn approaches 2 for higher polymers. Taylor8 has shown that the ratio is very nearly 2 for Gtinylon polyamide, and that the value of *Ifvis close to that of Mw. The weight average can therefore be used interchangeably with without introducing serious error. When mixtures of condensation polyesters or polyamides are considered, the importance of considering a weight average rather than a number average in calculating viscosity is apparent. The viscosity in the molten state of a mixture of polyesters was shown by Flory' to agree with the Mw, ( 6 ) Flory and Stickney. THISJ O U R N A L , 66, 372 (1!)&3), (7) ITlory. i h i J . , 62, 10.v~7,3032 ( l ! l l O ) . fa) 'L';iylor, ibtil., 6 9 , I j X , , 038 (l!)-l7),
